Use of computer image analysis for in vivo estimates of the carcass quality of bulls
The aims of the paper were to construct models for the estimation of carcass quality by means of computer image analysis and to verify computer photometry as an <i>in vivo</i> method of carcass quality prediction. Results of photometric measurements and carcass quality of 118 Slovak Pied bulls slaughtered at the age of 15 to 18 months were analysed. Nine length dimensions and four area dimensions were measured on the images of the top, left and rear view of each animal. Hot carcass weight (HCW), weight of meat in carcass (WMC) and weight of meat in valuable cuts (WMVC) were obtained after slaughter treatment and carcass dissection. HCW, WMC and WMVC revealed a maximum correlation with the top-view body area (<i>r</i> = 0.54–0.60) and thurl width (<i>r</i> = 0.58–0.60). Stepwise regression was applied to construct linear regression equations for HCW, WMC and WMVC in two alternatives using photometrical dimensions with and without weight before slaughter (WBS). <i>R</i><sup>2</sup> in an alternative without WBS were lower (<i>R</i><sup>2</sup> = 0.47–0.55); however <i>R</i><sup>2</sup> in an alternative with weight before slaughter were higher and highly significant (<i>R</i><sup>2</sup> = 0.83–0.92). In both alternatives, the equation for HCW had the highest <i>R</i><sup>2</sup> and the equation for WMVC had the lowest <i>R</i><sup>2</sup>. Equations using photometric dimensions and WBS are suitable to estimate HCW, WMC and WMVC without detailed dissection.